The molecular interactions implicated in the mammalian G1/S cell cycle phase transition comprise a highly non-linear network which can produce seemingly paradoxical results and make intuitive interpretations unreliable. A new approach to this problem was explored, consisting of (1) a convention of unambiguous reaction diagrams, (2) a convenient computer simulation method, and (3) a quasi-evolutionary method of probing the functional capabilities of components of the network (Kohn, Oncogene, in press). A novel aspect of the computational method is that the reaction diagram defines a computer-readable reaction file that represents the system at a "microworld" level, and it is not necessary to write out the differential equations explicitly. Simulations were carried out for a sequence of hypothetical primordial systems, beginning with the simplest plausibly functional system. The complexity of the system was increased in small steps, such that functionality was added at each step. A new functional concept emerging from this study was that Rb-family proteins could store E2F in a manner analogous to the way a condenser stores electric charge, and, upon phosphorylation, release a large wave of active E2F. Moreover, excessive or premature cyclin- dependent kinase activities could paradoxically impair E2F activity during the G1/S transition period. The results suggest that network simulations, carried out by means of the methods developed, can assist in the design and interpretation of experiments probing the control of the G1/S phase transition. Collaborations are being sought to apply theory to experiment. (In a related project, with D.S. Dimitrov as P.I., we have carried out similar explorations using a differential equations solver.)